Date: Wed, 29 Jun 2011 07:51:02 +0200
From: Joerg Roedel <joro@8bytes.org>
To: KyongHo Cho <pullip.cho@samsung.com>
Cc: Joerg Roedel <joerg.roedel@amd.com>, Ohad Ben-Cohen <ohad@wizery.com>,
        kvm@vger.kernel.org, linux-kernel@vger.kernel.org,
        iommu@lists.linux-foundation.org, Avi Kivity <avi@redhat.com>,
        David Brown <davidb@codeaurora.org>,
        David Woodhouse <dwmw2@infradead.org>
Subject: Re: [PATCH 0/2] Introduce iommu_commit() function
Message-ID: <20110629055102.GG29299@8bytes.org>
References: <1308843083-10442-1-git-send-email-joerg.roedel@amd.com> <BANLkTi=mfMfVVOmoZd49fqFnY_kq0hxzDw@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <BANLkTi=mfMfVVOmoZd49fqFnY_kq0hxzDw@mail.gmail.com>
User-Agent: Mutt/1.5.18 (2008-05-17)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1079
Lines: 29

On Wed, Jun 29, 2011 at 02:02:02PM +0900, KyongHo Cho wrote:
> Hi.
> 
> On Fri, Jun 24, 2011 at 12:31 AM, Joerg Roedel <joerg.roedel@amd.com> wrote:
> > It needs to be called whenever a some code changed a domain (either by
> > attaching/detaching devices or by mapping/unmapping pages in
> > the domain).
> 
> Do you mean we can invalidate IOTLB with this iommu_commit()?
> We need to invalidate IOTLB without updating page table for some
> optimized solutions.
> 
> If a device in one domain is moved to other domain, IOTLB of the
> device must be invalidated
> because it contains translation information of the previous domain.

The classic use-case is to do a single iommu-tlb flush after a set of
page-table updates, but flushing iotlbs after moving devices is a
potential use-case too, so the answer is yes.

Regards,

	Joerg

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/